Acute myeloid leukemia (AML) is a heterogeneous disease with diverse gene mutations and chromosomal abnormalities. Core binding factor (CBF) leukemias, those with translocations or inversions that affect transcription factor genes RUNX1 or CBFB, account for approximately 24% of adult acute myeloid leukemia (AML) and 25% of pediatric acute lymphocytic leukemia. The encoded proteins, RUNX1 and CBFbeta, form a heterodimer to regulate gene expression, and they are both required for hematopoiesis in vertebrate animals from zebrafish to man. Extensive clinical studies have demonstrated that CBFB-MYH11 and RUNX1-ETO, the two common fusion genes in CBF leukemia, are the best biomarkers for diagnosis, prognosis, and residual disease monitoring of CBF leukemia patients. Even though CBF leukemias have better initial remission rate and better prognosis than most AML cases, current chemotherapy is associated with significant morbidity and mortality, and the long-term survival (>5 year) is only around 50-60%. Over the years we have used mouse models and a variety of research tools to characterize the CBFB-MYH11 fusion gene, determine the effect of the encoded protein, CBFbeta-SMMHC, on normal hematopoiesis, and understand the leukemogenesis process associated with the fusion gene. We have generated both conventional and conditional knock-in mouse models to study CBFB-MYH11. Using these mouse models we have demonstrated that Cbfb-MYH11 dominantly inhibits Runx1 and Cbfb function during definitive hematopoiesis, resulting in complete lack of definitive hematopoiesis in the heterozygous Cbfb-MYH11 knockin embryos. We also showed that Cbfb-MYH11 is necessary but not sufficient for leukemia, and we were able to identify cooperating genetic events in the mouse models. We have generated knock-in mouse models expressing truncated Cbfb-MYH11 to determine the importance of functional domains of CBFbeta-SMMHC. Overall our lab has been recognized in the field as the major contributor to the understanding of CBFB-MYH11 leukemia. In the last fiscal year we focused on three specific aims in this project. In the first aim we generate mouse models to study the mechanism of leukemogenesis by CBFB-MYH11. We were particularly interested in the contributions of the C-terminal domains of the CBFbeta-SMMHC fusion protein to leukemogenesis. In our recent publication last year (Kamikubo et al., Blood 121:638, 2013) we demonstrated that the C-terminus of CBFbeta-SMMHC is essential to induce embryonic hematopoietic defects and leukemogenesis. This result suggests that the C-terminal region can be considered for future development of targeted therapy for this type of leukemia. However, the C-terminus has two functional domains, one responsible for multimerization of the fusion protein and the other for RUNX1 repression. In the last fiscal year we have generated a new mouse model to determine the functional importance of the multimerization domain of CBFbeta-SMMHC. Preliminary data suggest that this domain is indeed critical for hematopoietic defects and leukemogenesis. In the second aim we continued our development and characterization of novel compounds for targeted therapy against CBF leukemias. Our strategy is based on the hypothesis that the RUNX1-CBFbeta interaction is critically important for CBF leukemia, and inhibition of this interaction can be therapeutically beneficial to the CBF leukemia patients. The novel compounds that we are developing that target RUNX1-CBFbeta interaction will greatly enhance our ability to treat CBF leukemia patients with greater efficacy and fewer side effects than conventional chemotherapies, if we achieve our goals. In addition to CBF leukemia, recent collaborations with several groups indicate that we can broaden the potential applications of these compounds to other leukemia, and even some solid tumors. In FY12 we identified benzodiazepine compounds that bind to both RUNX1 and CBFbeta and inhibit their functions. In the last fiscal year we have continued our pre-clinical studies. In collaboration with the TRND program of NCATS, we have refined our mouse CBF leukemia model and confirmed the efficacy of one of the compounds for inhibiting CBF leukemia in this model, which was performed in an independent lab. We have also developed several in vitro models for testing the interactions between chemical compounds and RUNX1/CBFbeta, which will be useful for mechanistic studies and for testing additional chemical compounds. The final aim of this project is to characterize relapsed CBF leukemia through sequencing. Relapse occurs in 30-70% of the patients with acute myeloid leukemia, and is a major reason for patient mortality. This is especially true for CBF leukemia, since the initial remission rate is very high. Treatment options for relapsed leukemia are limited, and the only potentially curative therapy for relapsed leukemia is hematopoietic stem cell transplantation, which is not available for every patient. It has been shown recently that mutations or genomic rearrangements taking place in a subset of leukemia cells can lead to relapse. If we can identify such mutations in CBF leukemias through genomic sequencing studies, we will be able to better understand the molecular mechanisms underlying relapse and design more effective treatments for such patients. To pursue this goal, we applied genomic approaches (whole exome sequencing and single nucleotide polymorphism arrays) on DNA from samples collected at sequential time points (i.e., diagnosis, complete remission and relapse) in seven patients with CBFB-MYH11 and six patients with RUNX1-ETO. We identified mutations in several previously identified AML driver genes, and we found a common chromosome rearrangement associated with relapse. Furthermore, our data has identified previously unreported putative driver genes for CBF AML. Overall, our data revealed two distinct profiles that support different mechanisms of relapse: 1) diagnosis and relapse blasts harbor the same driver gene mutations, indicating the intrinsic resistance of the major clones present at diagnosis to treatment regimen used; 2) diagnosis and relapse tumors have different driver gene mutations, indicating disease clonal evolution possibly through treatment selective pressure.